U.S. patent application number 13/304201 was filed with the patent office on 2013-05-23 for wireless sensor network with energy efficient protocols.
This patent application is currently assigned to KING FAHD UNIVERSITY OF PETROLEUM AND MINERALS. The applicant listed for this patent is MUNIR A. KULAIB AL-ABSI, AHMAR SHAFI, FAROOQ SULTAN, SALAM A. ZUMMO. Invention is credited to MUNIR A. KULAIB AL-ABSI, AHMAR SHAFI, FAROOQ SULTAN, SALAM A. ZUMMO.
Application Number | 20130128786 13/304201 |
Document ID | / |
Family ID | 48426864 |
Filed Date | 2013-05-23 |
United States Patent
Application |
20130128786 |
Kind Code |
A1 |
SULTAN; FAROOQ ; et
al. |
May 23, 2013 |
WIRELESS SENSOR NETWORK WITH ENERGY EFFICIENT PROTOCOLS
Abstract
The wireless sensor network with energy efficient protocols
includes a network of external sensors in communication with a data
sink. The network utilizes an algorithm integrating a modified
S-MAC (an algorithm for medium access control) protocol for
decreasing energy usage in operating the node and associated
sensors. A routing protocol is further integrated into the
algorithm, the routing protocol being based upon cluster head
rotation.
Inventors: |
SULTAN; FAROOQ; (DHAHRAN,
SA) ; ZUMMO; SALAM A.; (DHAHRAN, SA) ;
AL-ABSI; MUNIR A. KULAIB; (DHAHRAN, SA) ; SHAFI;
AHMAR; (DHAHRAN, SA) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
SULTAN; FAROOQ
ZUMMO; SALAM A.
AL-ABSI; MUNIR A. KULAIB
SHAFI; AHMAR |
DHAHRAN
DHAHRAN
DHAHRAN
DHAHRAN |
|
SA
SA
SA
SA |
|
|
Assignee: |
KING FAHD UNIVERSITY OF PETROLEUM
AND MINERALS
DHAHRAN
SA
|
Family ID: |
48426864 |
Appl. No.: |
13/304201 |
Filed: |
November 23, 2011 |
Current U.S.
Class: |
370/311 |
Current CPC
Class: |
H04W 52/0238 20130101;
Y02D 70/22 20180101; Y02D 70/142 20180101; Y02D 70/144 20180101;
Y02D 70/162 20180101; Y02D 30/70 20200801 |
Class at
Publication: |
370/311 |
International
Class: |
H04W 52/02 20090101
H04W052/02 |
Claims
1. A wireless sensor network with energy efficient protocols,
comprising: a plurality of wireless nodes; means for forming at
least one fixed cluster from the plurality of wireless nodes, the
at least one fixed cluster having a cluster head and at least one
member node in operable communication with the cluster head; a sink
node; and means for establishing wireless communication between the
sink node and the at least one fixed cluster; wherein the sink node
communicates with the cluster head to obtain sensor data reported
by the at least one member node.
2. The wireless sensor network according to claim 1, further
comprising: a single master clock in operable communication with
said sink node; and means for unitarily scheduling all said nodes
of the wireless sensor network based on the single master
clock.
3. The wireless sensor network according to claim 2, further
comprising means for periodically swapping roles between said
cluster head and said at least one member node in a non-random
manner, wherein said at least one member node becomes said cluster
head and said cluster head becomes said at least one member
node.
4. The wireless sensor network according to claim 3, further
comprising means for forming a logical network tree, said at least
one fixed cluster having a tree depth based on a signal strength of
a beacon sent from said sink node to said at least one fixed
cluster.
5. The wireless sensor network according to claim 4, wherein said
means for establishing wireless communication between said sink
node and said at least one fixed cluster further comprises means
for assuring said cluster head transmits a flow of data directed
unidirectionally up said logical network tree in a direction
towards and ultimately destined for said sink node.
6. The wireless sensor network according to claim 5, wherein said
cluster head has a communications routing table, the network
further comprising means for updating the communications routing
table.
7. The wireless sensor network according to claim 6, further
comprising means for node sleep schedule synchronization based on
the beacon sent from said sink node to said at least one fixed
cluster.
8. The wireless sensor network according to claim 7, further
comprising means for programming transmission power to a lowest
setting for communications between said at least one member node
and said cluster head.
9. The wireless sensor network according to claim 8, further
comprising means for programming transmission power to a higher
setting for inter-cluster head communications.
10. A routing protocol method in a wireless sensor network having a
plurality of clusters, each of the clusters having a cluster head,
the network further having a sink node, the method comprising the
steps of waking up a first node of the wireless sensor network,
thereby energizing the first node's systems; having the first node
listen for a first keep alive data packet on an assigned
communications channel for a predetermined time period;
constituting the first node as a first cluster head if the
predetermined time period has timed out without the first node
receiving the first keep alive data packet; to after constituting
the first node as a first cluster head, having the first cluster
head transmit a second keep alive data packet; having the first
node measure a received signal strength indication (RSSI) if the
first node received the first keep alive data packet; having the
first node compare the RSSI to a threshold value; constituting the
first node as a first cluster head if the RSSI is below the
threshold value; constituting the first node as a first member node
if the RSSI is above the threshold value; and having the first node
follow scheduling from a single master clock of the wireless sensor
network.
11. The routing protocol method according to claim 10, further
comprising having said first cluster head perform the steps of;
receiving data packets from a second member node; serially
forwarding the data packets between the cluster heads in the
network for retransmission to said sink node if and only if
receiving cluster heads are topologically closer to said sink node
than the forwarding cluster heads in a multi-hop transmission
scheme until said sink node receives said data packets.
12. The routing protocol method according to claim 11, further
comprising the step of having said first cluster head periodically
switch roles with said second member node in a non-random manner,
whereby said second member node becomes said first cluster head and
said first cluster head becomes said second member node.
13. The routing protocol method according to claim 11, wherein when
a ready-to-transmit UGN (Unsynchronized Group of Nodes) node
receives a keep alive transmission from one of the cluster heads,
having the ready-to-transmit UGN node perform the step of
transmitting its UGN node data to the cluster head that transmitted
the keep alive transmission.
14. The routing protocol method according to claim 13, further
comprising the step of having the cluster head receiving said UGN
node data forward said UGN node data to said sink node according to
said multi-hop transmission scheme.
15. A wireless sensor network MAC protocol method in a wireless
sensor network having a plurality of clusters, each of the clusters
having a cluster head, the network further having a sink node, the
method comprising the steps of: having a ready-to-transmit sensor
data node of the wireless sensor network wait a first time until an
assigned communications channel used in the wireless sensor network
is free; having the ready-to-transmit sensor data node transmit a
Request-To-Send (RTS) packet on the assigned communications channel
to a receiving cluster head; having the ready-to-transmit sensor
data node wait until the ready-to-transmit sensor data node has
received a Clear-To-Send (CTS) from the receiving cluster head;
having the ready-to-transmit sensor data node of the wireless
sensor network wait a second time until the assigned communications
channel used in the wireless sensor network is free; having the
ready-to-transmit sensor data node of the wireless sensor network
transmit the sensor data to the receiving cluster head; and having
the read-to-transmit sensor data node sleep after transmitting the
sensor data.
16. The wireless sensor network MAC protocol method according to
claim 15, wherein said ready-to-transmit sensor data node is a
member node.
17. The wireless sensor network MAC protocol method according to
claim 15, wherein said ready-to-transmit sensor data node is a
cluster head node.
18. The wireless sensor network MAC protocol method according to
claim 15, wherein said RTS and CTS packets are Unicast packets.
19. The wireless sensor network MAC protocol method according to
claim 15, further comprising the step of having said receiving
cluster head send an acknowledgement (ACK) packet to said
ready-to-transmit sensor data node to indicate that said sensor
data was successfully received.
20. The wireless sensor network MAC protocol method according to
claim 15, further comprising the step of having said receiving
cluster head send said CTS only if there is sufficient time
available in a wake period to complete transmission and
acknowledgment of said sensor data.
Description
BACKGROUND OF THE INVENTION
[0001] 1. Field of the Invention
[0002] The present invention relates to wireless protocols and
networks, and particularly to a wireless sensor network with energy
efficient protocols.
[0003] 2. Description of the Related Art
[0004] To ensure proper operation of a wireless sensor network
(WSN), efficient power consumption must be ensured throughout the
network. In addition, self-healing ability of the network, cost
efficiency and flexible network architecture must also be
provisioned to enable seamless operation of the network. Power
consumption is the most important design factor for WSNs.
Conserving power at each node, eventually reads to the extension of
the overall network life. On the hardware front, efficient design
of the node could serve as a major factor in deciding the power
consumption figures. At the application level, power conservation
can be incorporated into the design of the protocols by introducing
novel design and implementation procedures that take the energy
reserves into account. For example, minimizing the number of
collisions or choosing the shortest path to the destination can
help save power.
[0005] A major part of the power is wasted during transmission and
reception of radio packets. Since transmission and reception is
inevitable, short distance transmission and simple circuitry for
modulation/demodulation can be employed to save power. It is
important to know the causes of energy wastage so that appropriate
action must be taken to overcome or reduce them. In WSNs energy
wastage occurs in three domains, namely, sensing, data processing
and communications. However, the losses during communications are
considered to be the major factor of the network life. The
different causes of energy wastage have been identified and
discussed in the following text.
[0006] Packet collisions cause nodes to retransmit, which, in turn,
results in wastage of the battery power. A collision occurs when
multiple nodes transmit at the same time. Since all the nodes share
the same channel, collision avoidance must be ensured for proper
delivery of packets within the network. The situation becomes even
worse when multiple packets start to arrive at a receiving node
simultaneously.
[0007] Overhearing means that a node starts receiving packets that
are not destined for it. In normal operation, a node receives a
packet and then starts to parse it. In this process, the node
determines the destination address in the packet header and
discards it if the receiving node is not the destination of the
packet. The time required to complete this process depends on the
length of the packet header and also on the location of the
destination address in the header.
[0008] The main objective of the WSN is to relay information to the
sink. Control packets are, however, necessary for establishing and
maintaining efficient performance of the system. A large number of
control packets decreases the effective data throughput of the
network, and also causes an increase in energy dissipation. So
there is a tradeoff between the number of control packets that need
to be sent and the throughput of the network. Ideally, control
packets should be sent when absolutely necessary to ensure that the
network is productive with respect to data packets.
[0009] The network must have a high level of fault tolerance in
order to be of any practical value. In the setup of a WSN, the
nodes are scattered so that the sensed parameters reported by
individual nodes represent the situation at different physical
locations. Usually individual nodes cannot directly communicate
with the sink node. In such cases, data must be relayed through
intermediate nodes until it reaches the sink. In case of failure of
multiple nodes in the network, the data should still reach the sink
node by re-routing or variable power adjustment methods. In short,
failure of individual nodes should not affect the operation of the
network.
[0010] WSNs may include hundreds if not thousands of sensor nodes.
New nodes may join the network and older nodes may die out without
informing the administrator. In such scenarios, the network must be
flexible enough to occupy the changes and to accommodate the
variable size while maintaining an acceptable level of
integrity.
[0011] The deployment cost is a very important design factor for
sensor networks because of the large number of nodes required, as
well as the fact that in most networks the nodes are disposable.
The cost includes both the hardware and the software required to
monitor the network.
[0012] The limited power supply of the nodes makes it inevitable to
use energy conservation techniques to design the network that lasts
a longer period of time. It is, therefore, important to know the
causes of energy wastage so that appropriate action must be taken
to overcome or reduce them.
[0013] Once the packets collide, the data gets corrupt. Such
packets have to be discarded, and retransmissions have to be
requested, increasing the energy consumption in the network.
Moreover, when the control packets collide, the complete network
setup is affected. The delay in packet delivery also increases due
to collisions. As an indirect consequence of retransmissions, the
effective throughput of the system decreases, since the majority of
the time is wasted in retransmissions.
[0014] Overhearing wastes valuable energy in reading and receiving
packets that are not intended for the desired node. Moreover, until
the completion of this process, all the other packets intended for
the node are not received, thus increasing the latency in the
network and resulting in collisions.
[0015] Over-emitting causes high power dissipation and must be
avoided by time synchronization schemes or scheduling. Although the
losses in idle listening are not severe, they must also be
minimized to increase the network lifetime.
[0016] Thus, a wireless sensor network with energy efficient
protocols solving the aforementioned problems is desired.
SUMMARY OF THE INVENTION
[0017] The wireless sensor network with energy efficient protocols
includes a wireless node using an off-the-shelf variety of
different sensors and components that can support them. The node
includes a high performance MCU (Microcontroller Unit) with a
double-sided design to reduce the size to a minimum. The use of
special stub antennas further decreases the overall height of the
node. On the software front, a MAC (Medium Access Control protocol)
layer protocol inspired from S-MAC (Sensor-MAC) has been modified
with some unique implementation procedures not mentioned in
literature. Useful features of both TEEN (Threshold sensitive
Energy Efficient sensor Network routing protocol) and LEACH (Low
Energy Adaptive Clustering Hierarchy routing protocol) have been
combined into a new protocol that employs a cluster head approach,
and at the same time is suitable for dynamic environments, in
addition to the incident-based Unsynchronized Group of Nodes (UGN).
Dynamic head shifting is employed to increase the lifetime of the
network. The system defines a total of eleven (11) cluster levels
so that the system can support large sizes of networks with small
multi-hop routed communications among the cluster heads (CHs), thus
saving power for transmission.
[0018] These and other features of the present invention will
become readily apparent upon further review of the following
specification and drawings.
BRIEF DESCRIPTION OF THE DRAWINGS
[0019] FIG. 1 is a block diagram of a wireless sensor node in a
wireless sensor network with energy efficient protocols according
to the present invention.
[0020] FIG. 2 is a perspective view of a transceiver module of the
wireless sensor network with energy efficient protocols according
to the present invention.
[0021] FIG. 3 is an elongate linear polarization antenna of the
wireless sensor network with energy efficient protocols according
to the present invention.
[0022] FIG. 4 is a block diagram showing the network topology of
the wireless sensor network with energy efficient protocols
according to the present invention.
[0023] FIG. 5 is a timing diagram of CH node and member node duty
cycles of the wireless sensor network with energy efficient
protocols according to the present invention.
[0024] FIG. 6 is a flowchart showing a sensor node bootup process
of the wireless sensor network with energy efficient protocols
according to the present invention.
[0025] FIG. 7 is a flowchart of member node channel contention
logic during data packet sending in a wireless sensor network with
energy efficient protocols according to the present invention.
[0026] FIG. 8 is a screen shot of an NMA management tool for a
wireless sensor network with energy efficient protocols according
to the present invention.
[0027] FIG. 9 is a diagram of a working NMA instance of the
wireless sensor network with energy efficient protocols according
to the present invention.
[0028] Similar reference characters denote corresponding features
consistently throughout the attached drawings.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0029] The wireless sensor network with energy efficient protocols
includes a wireless node using off-the-shelf sensors and components
that can support a plurality of different sensors. The node
includes a high performance microcontroller unit (MCU) with a
double-sided design to reduce the physical size to a minimum. The
use of special stub antennas further decreases the overall height
of the node.
[0030] At the outset, it will be understood that the diagrams in
the Figures depicting the wireless sensor network with energy
efficient protocols are exemplary only, and may be embodied in a
dedicated electronic device having a microprocessor,
microcontroller, digital signal processor, application specific
integrated circuit, field programmable gate array, any combination
of the aforementioned devices, or other device that combines the
functionality of the wireless sensor network with energy efficient
protocols onto a single chip or multiple chips programmed to carry
out the method steps described herein, or may be embodied in a
general purpose computer having the appropriate peripherals
attached thereto and software stored on a non-transitory computer
readable media that can be loaded into main memory and executed by
a processing unit to carry out the functionality of the apparatus
and steps of the method described herein.
[0031] The node elements group themselves into clusters and decide
a cluster head (CH) that communicates with its peers to route data.
The protocol includes (CH)s that are rotated periodically in a
pre-defined manner to facilitate energy dissipation of the nodes.
The nodes form fixed clusters, and they retain their membership
till the end. This not only simplifies the implementation, it also
results in a savings of energy from the overhead of communication
required for making new clusters. The MAC protocol described herein
assumes that all the nodes can receive signals from the sink, which
can become difficult in different environments. The preprogrammed
thresholds can become difficult to change as priorities for alarms
change.
[0032] The system settings can be modified very easily in the
system software at any time, depending upon the particular
requirements of the application for which the WSN is being
employed. All nodes are identical and act both as a sensing node
and as a cluster head (CH) on their turn. The network topology is
cluster-based. A data sink node broadcasts a beacon signal every
four seconds. At startup, every node waits for five seconds before
deciding whether it should declare itself as a new cluster head, or
whether it should join an existing cluster as a member.
[0033] Useful features of both TEEN and LEACH have been combined
into a new protocol that employs the aforementioned cluster head
approach, and at the same time is suitable for dynamic
environments, in addition to the incident-based Unsynchronized
Group of Nodes (UGN). Dynamic head shifting is employed to increase
the lifetime of the network. The system defines a total of eleven
(11) cluster levels, and thus can support large sizes of networks
with small multi-hop routed communications among the CHs, thus
saving substantial power for transmission.
[0034] The data acquisition part of a WSN comprises a large number
of tiny wireless devices called nodes, which are deployed over a
physical environment and actively cooperate in order to accomplish
one or more tasks.
[0035] As shown in FIG. 1, a typical sensor node 100 includes a
power supply 108, sensors 106, a microcontroller unit (MCU) 104,
and a transceiver 102 to send and receive data. The power supply
108 is used to power the node. The sensor circuitry can transform
physical quantities into an electric signal. An analog-to-digital
converter (ADC), typically a part of the MCU 104, changes the
analog signals generated by the sensors into digital signals and
sends them to the processing portion of subsystem 104. The
processor can then perform simple operations on the received
digital signal and can store it into memory. Finally, the
transceiver 102 sends and receives data to different destinations
as and when required. The sensor nodes relay their sensed data
through each other or directly to the base station, depending on
the scale of the network and their position with respect to the
base node.
[0036] The WSN protocols involve a cluster head approach, which
provides any given sensing node with only a single relay (cluster
head), so that the number of hops between the sensing node and the
destination are decreased, thereby effectively decreasing overall
packet latency. The base station may send control commands
(downstream messages) down into the networks, for example, a
request to increase their sampling frequency. Sensors are designed
to support unattended operation for long durations, frequently in
remote areas, in smart buildings, or even in hostile environments.
A set of sensing components forms a part of the device. Popular
examples of sensing components include temperature sensors,
accelerometers, humidity sensors, infrared light sensors, pressure
sensors, and magnetic sensors, as well as chemical sensors. WSNs
have applications in all fields, such as structural monitoring,
environmental monitoring, and object tracking. Owing to the
hostility and remoteness of the operating region, some critical
factors for the efficient operation of a WSN include power
efficiency for longer network life, fault tolerance for network
integrity, scalability for variable network topologies and dynamic
network setup, which are just some of the desirable features for a
practical WSN.
[0037] A data distribution part of the network is responsible for
collecting the data from the base station node and distributing it
through different media. WiFi, LAN and even long distance radio
communication techniques are used to accomplish this task. The work
done by the present invention is at the data acquisition level of
the WSN. Designs at hardware level (sensor node) as well as
protocol level are done to ensure cost as well as energy
efficiency.
[0038] The inventive protocols are aimed at increasing the network
operating time. To fulfill this requirement the selected medium
access control (MAC) and routing protocols must be energy efficient
and should have the ability to provide an acceptable data
throughput for the network to be meaningful.
[0039] The routing protocols include cluster-based hierarchical
protocols that have proved to provide a much better data throughput
for large networks having hundreds of nodes. The issue of selecting
a permanent cluster head leads to higher energy dissipation and
therefore a reduced network life; since the death of the cluster
head causes issues in the selection of the new cluster head. When
the cluster head runs out of power, until the selection of the new
cluster head, the cluster remains out of touch from the rest of the
network and in doing so leads to multiple packet loss. The system
streamlines the head selection procedure so that the network is not
disrupted while providing an acceptable level of end-to-end
throughput.
[0040] The implementation of the designed protocol on hardware
presents multiple challenges. Although the design details are
presented in detail in literature, no mention of implementation
process is done. Due to the unavailability of implementation
details, unique methods must be devised while ensuring the network
operation is not compromised.
[0041] In order to deploy a WSN in a given area, the least number
of nodes required to cover a major part of the area must be
determined. Since the number of nodes is directly related to the
cost of the network, the process becomes a crucial' part of the
network deployment stage.
[0042] WSN can be deployed in buildings to monitor live activities
like temperature/humidity and in turn operate the heating/cooling
equipment depending upon the sensed values. In addition, intrusion
detection can be accomplished to enhance the security situation of
a locality. Intelligent buildings employing systems like BACnet are
examples of this approach.
[0043] WSNs have been used to monitor environmental phenomenon such
as, e.g., volcanic activity to alert evacuation teams beforehand.
This type of system has also been used to monitor
temperature/humidity condition in tea plantations to operate the
irrigation system autonomously.
[0044] WSNs in the form of body area networks (BANs) have been used
to monitor health activity of patients remotely. The network
gathers the vital information and sends it to the doctor over the
Internet. Thus, the doctor has the complete up to date information
about the patient without any dedicated monitoring.
[0045] An existing medium access protocol called the sensor medium
access control (S-MAC) has been used for ad-hoc WSNs. This protocol
depends on the request-to-send (RTS)/clear-to-send (CTS) mechanism
of the IEEE 802.11 to avoid collisions. It was shown that S-MAC has
2-6 times less power consumption than IEEE 802.11. Timeout MAC
(TMAC) protocol exists and has been proposed to solve the problem
of idle listening in a wireless sensor network. The T-MAC
dynamically adapts a listen/sleep duty cycle through finely grained
timeouts, while having minimum complexity. Both S-MAC and T-MAC
operate on the principle of adopting sleep-wake schedules to
synchronize the wake timings of all the nodes within a
vicinity.
[0046] To further reduce the energy wastage due to idle listening,
low power listening (L.PL.) has been employed to reduce the duty
cycles to less than 0.1%. WiseMAC and B-MAC are examples of an LPL
implementation. In LPL nodes wake up for a very brief period to
check the channel activity without actually receiving data. This is
called channel polling. There also exists a Power control mechanism
based on S-MAC protocol and modifies the competitive mechanism of
S-MAC resulting in enhanced S-MAC (ET-MAC). The energy-efficient
and high throughput MAC (ET-MAC) protocol embeds some extra
information in long wake up preamble frame and it also uses
collision avoidance signaling and handshaking. These ideas help
wireless nodes to stay at sleep node as much as possible. Their
simulation results and analysis showed that with dynamic traffic
load, this protocol achieves improvement in energy-efficiency.
[0047] For multiple access control collisions are avoided by
virtual and physical carrier sense (CS). There is a duration field
in each transmitted frame that indicates how long the remaining
transmission will be for. If a node receives a packet destined to
another node, it knows how long to keep silent from this field. The
node records this value in a variable called the network allocation
vector (NAV) and sets a timer for it. Every time when the timer
fires, the node decrements its NAV until it reaches zero. Before
initiating a transmission, a node first looks at its NAV. If its
value is not zero, the node determines that the medium is still
busy. This is called virtual carrier sense. Physical carrier sense
is performed at the physical layer by listening to the channel for
possible transmissions. Carrier sensing time is randomized within a
contention window to avoid collisions and starvation. The medium is
determined as free if both virtual and physical carrier sense
indicates that it is free. All sender nodes perform carrier sense
before initiating a transmission. If a node fails to get the
medium, it goes to sleep and wakes up when the receiver is free and
listening again. Unicast packets follow the sequence of
RTS/CTS/DATA/ACK between the sender and the receiver. After the
successful exchange of RTS and CTS, the two nodes will use their
normal' sleep time for data packet transmission. In order to
explain this protocol some assumptions have been considered. This
protocol assumes that all the nodes in the network are stationary
and are able to transmit at variable levels of power. The nodes are
deployed randomly and have the ability to communicate with
neighboring nodes. To conserve energy, most of the data
communications is done between neighbors rather than between end
points. This means, to transmit data between two locations, nodes
transmit to their neighbors and the data reaches the destination
over a multi-hop path. The nodes are assumed to operate dedicated
applications that result in the activity being reported. This
enables the nodes to be hard-coded rather than generally
programmed. Every node has the ability to aggregate redundant data
before transmission to avoid congesting the network by sending
copies of the same packet. The applications are assumed to tolerate
some level of latency since environmental monitoring will have long
idle periods followed by data bursts in case of alarm.
[0048] Threshold sensitive Energy Efficient sensor Network protocol
(TEEN) is a hierarchical protocol' designed to be responsive to
sudden changes in the sensed environmental attributes such as
temperature. Responsiveness is important for time critical
applications, in which the network operates in a reactive mode.
[0049] The network architecture in TEEN is based on a hierarchical'
grouping where closer nodes form clusters and this process goes on
the second level until base station (sink) is reached. After the
clusters are formed, the cluster head broadcasts two thresholds to
the nodes. These are hard and soft thresholds for sensed
attributes. The hard threshold is the minimum possible value of an
attribute to trigger a sensor node to switch on its transmitter and
transmit to the cluster head. Thus, the hard threshold allows the
nodes to transmit only when the sensed attribute is in the range of
interest, thus reducing the number of transmissions significantly.
Once a node senses a value at or beyond the hard threshold, it
transmits data only when the value of that attribute changes by an
amount equal to or greater than the soft threshold]d. As a
consequence, soft threshold will further reduce the number of
transmissions if there is little or no change in the value of
sensed attribute.
[0050] This model uses hierarchical clustering scheme in which the
nodes are classified into clusters (exemplary clusters 402a, 402b
402c, and 402d are shown in FIG. 4). Nodes are grouped into
clusters and a cluster head is selected from amongst them. The
selection of cluster head can be fixed but is usually taken as
dynamic to distribute the energy consumption evenly throughout the
network. Once the cluster head has been selected, the remaining
nodes become its members. All the member nodes sense and forward
the data to their respective cluster heads. Once the head node gets
the data, it needs to forward it to the next head in the upper
level. This process continues until the packet reaches the
sink.
[0051] Hard threshold (HT) and soft threshold (ST) are defined by
the user and serve as the limits of the forwarded data. HT defines
the limit for the minimum sensed value after which the data should
be transmitted, whereas ST defines the minimum difference between
the sensed value and HT. By using HT and ST together, excessive
transmissions are avoided.
[0052] LEACH belongs to the proactive family of routing protocols.
The LEACH protocol enables all the nodes to sense at regular
intervals, turn on the radios, transmit the information and go back
to sleep. In other words, a snapshot of the sensed values is
provided after a particular time.
[0053] Since sensor nodes may die randomly, LEACH employs the
concept of dynamic clustering where the cluster memberships keep
varying from time to time. LEACH is completely distributed and
requires no global knowledge of the network, however, LEACH uses
single-hop routing where each node can transmit directly to the CH
and the sink. Therefore, it is not applicable to networks deployed
in large regions. Furthermore, the idea of dynamic clustering
brings extra overhead, e.g. head changes, advertisements etc.,
which may diminish the gain in energy consumption.
[0054] When the network is set up, the nodes are distributed into
clusters. Once the clusters have been created, the head node
transmits TDMA-based schedules to notify the member nodes of their
respective time slots during which communication takes place. When
all the nodes have sent the data to the cluster head, the head node
aggregates the data and forwards to the cluster head in the second
level. Due to time slots, the cluster heads has to be awake for a
larger amount of time as compared to the member nodes. Also the
aggregation of data requires processing, which in turn consumes
valuable energy.
[0055] Due to the high energy dissipation at the cluster head, it
is expected to die out early. To prevent this event from happening,
the cluster head responsibility is rotated among the member nodes.
This enables uniform energy consumption in the network ensuring
that no single node dies out earlier.
[0056] Traditional WSN design focuses on improving energy
efficiency and data throughput by adopting design methods for MAC
and routing protocols, however, the energy consumed by the wireless
node during the operation of the application is usually side lined.
The approach adopted for the design and implementation of the MAC
and routing protocols ensures energy efficiency on the application
level by using programming platforms like TinyOS that insure a high
degree of energy efficiency by limiting the resources being
used.
[0057] The protocols have been simulated and fine-tuned before
implementing them on actual hardware. The network has been tested
with 30 nodes for a sufficient period of time to confirm the
operation of the programmed protocols. In order to create a model
to provide an approximation on the network size, received radio
power measurements have been performed at two different locations.
This data has been used to construct a model that can provide the
approximate network size for a given region of interest.
[0058] The present system is a one-of-a-kind wireless sensor node
that can be programmed by a cheap off-the-shelf programmer rather
than using special sensor boards (MIB 510) as for the Crossbow Mica
devices. Temperature and light sensors have been included on the
node and extra ports have been provided to interface additional
sensors if required.
[0059] A MAC layer protocol inspired from S-MAC has been
implemented with some unique implementation procedures not
mentioned in literature. LEACH and TEEN have been used as
references to devise a routing protocol that serves the purpose of
data routing as well as keeps the power consumption in check. An
interactive network monitoring application has been created to
observe the network conditions and to record the data for future
usage. A network deployment application has been completed that
estimates the approximate network size based on the environment and
the target area.
[0060] The complete system has been tested with twenty-five nodes
with several clusters at different levels from the sink node. The
network performed well and remained stable throughout the
experiment.
[0061] The wireless sensor network with energy efficient protocols
utilizes a novel energy model that specifies activity-sleep
requirements for node components.
[0062] The wireless sensor network with energy efficient protocols
utilizes a highly power-efficient TinyOS operating system for
application design. Moreover, the wireless sensor network with
energy efficient protocols utilizes an AT megal28 supporting
TinyOS. The ATmegal28 is a low-power, high-performance 8-bit
microcontroller offering 128 KB of in-system self-programmable
flash program memory, 4 KB EEPROM, and 4 KB internal SRAM. It
allows executing powerful instructions in a single dock cycle,
achieving throughput approaching 1 million instructions per second
(MIPS) allowing the system designer to optimize power consumption
versus processing speed. It comes in surface-mount compact fool'
factor, thus achieving smaller size circuit designs.
[0063] A mandatory requirement of the WSN design is to conserve the
power in the sensor nodes. Hence, this microcontroller has support
for different power save modes. Sleep modes enable the application
to shut down unused modules of the MCU, thereby saving power. The
ATmegal28 provides six different power-save modes which includes an
Idle Mode wherein the CPU stops but the SPI, analog comparator,
ADC, two wire interface, timers/counters, watchdog and the
interrupt system continue to operate normally. This mode also
enables the MCU to wakeup from an external trigger.
[0064] An ADC Noise Reduction mode turns off the CPU but keeps the
ADC, two-wire interface, interrupt counter and timer/counterO keep
operating. This mode reduces the noise during ADC operation and
also enables higher resolution measurements.
[0065] A Power-down mode turns off the CPU and the external
oscillator while keeping the two-wire interface, external
interrupts and the watchdog counter operating. This mode reduces
the noise during ADC operation and also enables higher resolution
measurements. The MCU can wakeup from this mode only from external
interrupts.
[0066] A Power-save Mode is the same as compared to power down mode
except that all the timers are also enabled. This enables the MCU
to wakeup when timers/counters overflow.
[0067] A Standby Mode that is different from the power-down mode
allows the oscillator to continue to run. This feature can be used
when the oscillator is being used to time external circuitry
attached to the MCU.
[0068] An Extended Standby Mode is similar to the power save mode
with the exception that the oscillator keeps operational.
[0069] The wireless sensor network with energy efficient protocols
utilizes nodes that require several buffers and require keeping
track of different timing events. Preferably 32 general-purpose
working registers, and a real-time counter (RTC), as well as four
flexible timers/counters with compare modes are employed by the
wireless sensor network with energy efficient protocols. The
sensors preferably utilize 10-bit ADC with optional differential
input stage with programmable gain.
[0070] For RF subsystem, a CC2420 transceiver module 102 is used,
as shown in FIG. 2. The component CC2420 is a 2.4 GHz radio
frequency (HF) transceiver chip for low voltage, low power wireless
applications, which is used to interface low cost microcontrollers
and has very useful features such as the compact size as well as
data buffering, encryption and authentication. It provides the
capability of controlling the transmission power levels so as to
meet the requirements of different scenarios of the WSN
applications. Moreover, the transceiver module 102 includes a
compact antenna 202
[0071] In addition, CC2420 provides burst transmissions, channel
assessment, link quality indication and packet timing information.
These features reduce the load in any host microcontroller. The
CC2420 uses serial peripheral interface (SPI) to communicate with
the microcontroller four wires serial. Additionally there are four
extra lines out of the CC2420 to assist the microcontroller in
monitoring the status of the wireless communication. It has 256
Bytes buffer for transmitting and receiving data. The CC2420
supports ZigBee with direct sequence spread spectrum (DSSS)
transmission at 2 Mchips/sec and provides an effective data rate of
250 kbps. Moreover, CC2420 has the ability to adjust the
transmission power levels as indicated in Table 1.
TABLE-US-00001 TABLE 1 Transmission Power Levels Power Level
Transmitted Power in dBm 31 0 27 -1 23 -3 19 -5 15 -7 11 -10 7 -15
3 -25
[0072] The wireless sensor network with energy efficient protocols
can use a daughter board module CC2420EM. It is a Zigbee enabled RF
transceiver with TinyOS support produced by Chipcon. This board has
the CC2420 module with its configurations parts and circuits ready
and soldered on a PCB with a 40-pins socket along with its antenna
as shown in FIG. 2.
[0073] For the circuitry design all efforts have been made to
reduce the size and weight of the sensor nodes by designing more
efficient PCBs and using smaller components with better surface
space utilization e.g. double sided PCBs. A smaller size antenna is
preferable on the RF transceiver.
[0074] The wireless sensor network with energy efficient protocols
can be equipped with two types of sensors for environmental
monitoring, e.g., light and temperature. The light sensor may be
installed onboard and the temperature sensor may be interfaced with
the node via the expansion ports. Nodes sense the raw data and
provide the corresponding analog voltage to the microcontroller
that is converted to digital' format by the internal ADC of the
microcontroller for the processing and later transmission by the RF
transceiver to the neighboring nodes. Some of the features of the
inventive wireless sensor network node are that it has a reduced
area with double-sided PCB design structure. Moreover, the overall
design is improved in that the dimensions are reduced to a 6.09
cm.times.4.19 cm area. A smaller, more compact antenna is used to
reduce the effective height of the platform. In addition, the
optimization in physical area is achieved by relocating some
components and using comparatively compact passive components e.g.
by using small size 1/8 Watt resistors instead of the standard 1/4
Watt. The crystal oscillators are also used in a way so as to
conserve the space.
[0075] The LEDs used in the new design are surface mount type and
have smaller physical dimensions thus freeing up crucial space and
saving the area of the node. The actual area of the ultimate
functional node can be even smaller when the LEDs and their
peripheral circuitry are removed as they are used in the prototype
only for the debugging purposes to let us know the different states
of the node during its operation. The current assignments of the
LEDs include All OFF: Node in Sleep Mode, Green ON: Member Node
awake, Blue ON: Node sending data, Red ON: Head node awake
[0076] The wireless sensor network is a totally self-powered unit
and has an onboard battery pack as the power source. All the
wireless nodes in the network are identical in every aspect, except
the sink/base node, and they are powered using batteries. Only the
sink node is powered by commercial power via the USB port of the
Network Monitoring Server. Moreover, an ON-OFF switch is embedded
into the design to enable resetting the devices and to avoid the
wastage of energy during testing. The wireless nodes are powered by
3 AAA sized (1.5 VDC) batteries providing a total supply of 4.5V.
The wireless node has a set of extension ports that provides the
capability to add external' sensors. Currently, the node houses
onboard light sensor and has provision of installing temperature
and humidity sensors on board using those extension ports. The
wireless sensor network with energy efficient protocols has the
option of an additional analog sensor and a digital sensor with a
4.5V power port also available for powering the sensors.
[0077] In addition to the wireless sensor nodes, one sink node is
required to receive the data from the wireless network. This sink
node has one of the purposes of collecting the data from the sensor
nodes in the wireless network and forwarding it to a server so that
proper analysis and actions of the data can be performed.
Generally, it is assumed that the sink node has unlimited reserves
of energy as it is connected to the commercial power supply.
Moreover, the sink node is also used to synchronize all the nodes
in the network by sending a beacon signal at regular intervals.
[0078] The sink node connects serially to a PC USB port using a
serial-to-USB connection type converter. Connecting via serial port
requires the conversion of the local circuit voltages (known as TTL
level) to higher levels (R.S-232) so that serial ports can
understand the data being received. MAX232CPE by MAXIM Corporation
is an example of a voltage level converter. This IC when connected
along with peripheral circuitry converts the circuit level voltages
(=5V) to serial eve (=15V) and vice versa. This enables the
wireless sensor network with energy efficient protocols to connect
the base station with the PC through the serial port. Once the
connection is made, the serial port is read through software to
view the incoming packets.
[0079] In the wireless sensor network with energy efficient
protocols, each node plays a member role (of the cluster) and a
head role of the cluster. To save the power of the battery both
these types of nodes do not sense and transmit the data
continuously, instead, they wake up for a small period of time and
then sleep. Head nodes wake up for a little more time as compared
to member nodes. Head nodes also receive data from member nodes and
then transmit towards the sink node, whereas member nodes wake up
for lesser time and do not receive data from other nodes except CTS
byte from the head node and transmit data to the head node. The
nodes in head node mode consume more power than the member node
mode. In order to consume the same average power in all the nodes,
each node in the same cluster will become a head node at its own
turn. Moreover, the protocols described herein support the
integration of silent observer nodes, which do not broadcast any
sync packets and can only be used to demand data from any of the
nodes in the network. The interfacing of these observer nodes does
not require any additional programming and any observer node
connected to a terminal (a PC) can be deployed anywhere in the
network, provided it does not send any synchronization packets like
the sink node.
[0080] In order to set up the WSN, protocols must be implemented
both at the MAC as well as on the routing layer to first establish
the link level connections up to the sink node and then to route
the sensed data reliably. The primary objective in WSN protocol
design is to maximize the node and ultimately the network lifetime.
These protocols have to ensure communications network that is
robust, reliable, and flexible as well as energy
conservative/efficient at the same time. The protocols must also
ensure that the topology changes are well covered. The birth or
death of a node must not affect the operation of the network and
the data acquisition as well as the routing operations should
continue in a streamlined manner. For the sink node to receive the
data from all the nodes in the network, a routing protocol must be
implemented such that the data is reliably propagated through the
network. At the network layer, the main aim is to find ways for
energy-efficient route setup and reliable routing of the data from
the sensor nodes to the sink node such that the lifetime of the
network is maximized.
[0081] Routing in WSNs is very different from ordinary networks due
to several characteristics that distinguish them from contemporary
communication and wireless ad hoc networks, such as the
following.
[0082] In WSN it is not possible to build a global addressing
scheme for the deployment of large number of sensor nodes.
[0083] On the contrary to ordinary networks, almost all
applications of sensor networks require the flow of sensed data
from multiple regions (sources) to a particular sink.
[0084] The most important factor is that sensor nodes are tightly
constrained in terms of transmission power, on-board energy,
processing capacity and storage and thus require careful resource
management. Due to such differences, the present wireless sensor
network with energy efficient protocols employs new algorithms
solving the problem of routing data in sensor networks. The present
network, thus, utilizes the networking protocols discussed below
and relies on the following assumptions.
[0085] The network is assumed to be static, i.e., the nodes are
stationary and will remain in the same physical position throughout
the lifetime of the network. All the nodes have same energy
reserves, i.e., they are running on exactly the same type of
batteries. The nodes are placed/scattered randomly in the field and
the nodes form the network and later organize themselves
automatically. The sink node has unlimited power supply, therefore
power conservation techniques need not apply to the sink. The
coverage of the sink node is assumed to be all over the network. In
other words, all the nodes can hear the packets transmitted by the
sink. Only the member nodes are responsible for sensing. The CH
nodes do not perform sensing and serve for relaying and routing the
data packet from sensing nodes towards the sink node. The member
nodes communicate in Unicast only with their corresponding CHs. The
keep alive messages are not acknowledged. For demonstration
purposes and due to the limited number of nodes available, it is
assumed that all the nodes are grouped in maximum of 10 clusters
distributed in 5 levels (although the values of the power levels
can be any of the 11 allowed levels. We have also assumed that only
two clusters can lie at any single power level However, the
implemented networking protocols are totally flexible and can be
programmed to support much larger area and number of
nodes/clusters, if required.
[0086] The current settings of the system and network are as
follows. These settings can be modified very easily in the system
software any time depending upon the particular requirements of the
application for which the WSN is being employed. All nodes are
identical and act both as sensing node and act as CH on their turn.
The network topology is of cluster based. The sink node broadcast
the beacon signal at every 4 sec. At startup, every node waits for
five seconds before deciding whether it should declare itself as a
new cluster head or to join an existing cluster as a member. The
received signal' strength indicator (RSSI) value threshold is -5
dBm. The member nodes send keep-alive messages to their CH once
every cycle. The duration of the cycle of all the nodes in the
network is 4000 ms or four seconds. A cluster head remains ON for
300 ms and steeps for 3700 ins or 3.7 seconds. The member nodes are
awake for 100 ms and asleep for 3900 ms or 3.90 seconds, thus
keeping the same duration of the cycle as that of the cluster
heads. The responsibility of CH is shifted to other members of the
cluster after every 10 cycles or forty seconds. The cluster ID is
generated by the first CR by padding 0 to its own hard-coded node
ID. This cluster ID is maintained by all the nodes throughout the
cluster life. The communication between the member node and CHs
occurs at the lowest transmission power of the transceiver (mode 1)
whereas for inter-CH communications, a higher transmission power
level (mode 3) is used. If a CH does not receive a keep-alive from
a particular node for more than 3 cycles, it is considered dead and
is removed from the cluster table. After the rotation of the CH,
the routing tables are updated within five seconds. The data is
relayed from the sensing node to the sink node by multi-hop dynamic
routes. In the main synchronized network, the environment
attributes are being sensed continuously/periodically and being
relayed to the sink on a regular basis. In the unsynchronized
portion of the network, the nodes keep sensing the environment on
continuous basis but report the values only in case of a predefined
incident. Routing tables are updated after every twenty seconds.
The levels once assigned to a node will remain constant during the
lifetime of the node.
[0087] Plot 500 of FIG. 5 shows the complete timing process
graphically. The sink node sends the beacon periodically after
every four seconds. When a node wakes up it listens for five
seconds, receives the beacon and declares itself head by
broadcasting the keep-alive and sleeping after 300 ms. Similarly a
member node on boot-up receives the keep alive of the CH and goes
to sleep after 100 ms as indicated.
[0088] As mentioned above, the MAC protocol' that wireless sensor
network with energy efficient protocols actually uses has many
ideas from the S-MAC protocol. Several factors were not mentioned
in the literature and have been devised to accomplish the desired
tasks.
[0089] The choice to go for hierarchical-based protocol was made to
simplify the data routing process. If peer-based routing was
selected, then every sensor node had to know every other sensor
node in its surrounding causing memory wastage in updating and
maintaining node tables. In hierarchical-based protocols, only the
head maintains the address tables of its peers, and the member
nodes do not need to have any tables. During the implementation of
the MAC and network protocol we have made the following assumptions
that have to be considered while understanding the operation of our
network:
[0090] As suggested in S-MAC protocol, there are two types of
packets in the inventive implementation; namely, control packets
and data packets. The specific format of these packets has not been
defined in S-MAC and therefore the payload structure of each of
these packet types is defined depending on the intended
application's need and requirements. The packets are parsed into
two sections; namely, the header part and the payload part. The
header section of the packet is created by the transceiver and we
don't have control over it. Different fields for different
operations are defined in the payload section of the packet both in
the control packet and in the data packet.
[0091] Control packets are the packets that are responsible for
setting up the point-to-point links and clusters, establishing the
network topology and ensuring smooth and reliable data transfer
between the nodes. The control' packets are obvious]y overhead for
the network operation and ultimately decrease the effective
throughput of the system. Therefore, they are desired to be
minimally transmitted and only if/when required. Since there is no
data in these packets, preferably they should be made as small as
possible so that the least energy is wasted for the transmission of
this packet type. The format of the control packets is shown in
Table 2. It is a 16 bytes packet with the header of 13 bytes long
and the remaining 3 bytes are used in different scenarios for
different purposes as explained below.
TABLE-US-00002 TABLE 2 Control Packet Format Byte number Function
Type 1, 2 Packet start identifier Radio Header 3 Leading zeros
Radio Header 4, 5 Destination ID Radio Header 6, 7 Source ID Radio
Header 8 Payload Length Radio Header 9 Group ID Radio Header 10
Message type Radio Header 11 Cluster ID (CID) Radio Header 12 Node
level (NL) Radio Header 13 Identifier Radio Header 14, 15 Checksum
Radio Header 16 Packet end identifier Radio Header
[0092] Cluster ID (CID)(1 byte): This field in the packet
identifies the cluster ID of the sending node. Each cluster has a
unique ID assigned by its first cluster head (CH) at the time of
the birth of the new cluster and all the nodes belonging to that
cluster are identified by the cluster ID.
[0093] Level (L) (1 byte): This field identifies the hierarchical
position of the cluster in the routing tree. To enable routing of
the packets to the sink node, each cluster assigns itself a power
level. This is discussed in more detail in the routing protocol
implementation.
[0094] Type (T) (1 byte): This field acts as an identifier for the
type of the control packet. It distinguishes between the different
types of control packets that we have formulated. Table 3 gives the
details of all the control packet types, their use in the network
formation and maintenance and their corresponding values in the T
field.
[0095] Sensing Node ID (NID) (1 byte): This field carries the ID of
the sensing node. This unique ID is assigned to each node at
boot-up and is hard-coded inside every node. Each node is given a
unique ID that is not repeated in the remainder of the network.
Intermediate Node ID (IID1-IID5) (5 bytes): These 5 fields are used
to convey the complete route information the packet traverses from
the sensing node to the sink. In the current implementation, we
have programmed the system to operate for a maximum of six hops
including the hop from the sensing node to the CII. If the packet
reaches the sink with less than six hops, the remaining
corresponding fields are set to zero; otherwise, the fields have
the node IDs of all the intermediate heads that relay the data
packet. Sensing Node Level (1 byte): This field contains the
assigned power level of the sensing node.
[0096] Intermediate Node Level 1 (IL1-IL5) (5 bytes): These fields
provide the power levels of the intermediate CHs.
TABLE-US-00003 TABLE 3 Control Packet Type Type Field Argument
Packet Type Flow Direction 0x00 Keep alive within a CH to member
nodes cluster 0x01 Keep alive packet CH to other CHs between CHs
0x0A RTS Sending node to receiving node 0x0B CTS Receiving node to
sending node 0x0C CH control transfer Sent from current CH to
request the new CH 0x0D CH control transfer Sent from the new CH to
response current CH
[0097] The data packet is used to carry the data, from the sensing
member node to the corresponding CH, and subsequently from the CH
to CH until the data reaches the sink node and ultimately received
by the server. The complete length of the data packet is 31 bytes
including the header and the payload. The header is 13 bytes tong
and the remaining portion of the packet is the payload that we have
formulated according to our requirements.
[0098] We have used the payload part of the data packet not only to
carry the sensed data but also to carry the routing information
along the route (for the debugging purpose only) as well as to
display the routing information on the front-end application (for
the demonstration purposes only). In the practical' deployment of
the network, most part of this routing information is not needed
and thus the packet size can be reduced to just 4 bytes required
for storing light and temperature data. The format of the data
packet is shown in Table 4.
TABLE-US-00004 TABLE 4 Data Packet Format Byte number Function Type
11, 12 Temperature value (Temp) Payload 13, 14 Node level (NL)
Payload 15 Hop count (HC) Payload 16 Cluster ID (CID) Payload 17
Sensing node ID (NID) Payload 18 Intermediate node ID 1 (IID1)
Payload 19 Intermediate node ID 2 (IID2) Payload 20 Intermediate
node ID 3 (IID3) Payload 21 Intermediate node ID 4 (IID4) Payload
22 Intermediate node ID 5 (IID5) Payload 23 Sensing node level (SL)
Payload 24 Intermediate Node Level 1 (IL1) Payload 25 Intermediate
Node Level 2 (IL2) Payload 26 Intermediate Node Level 3 (IL3)
Payload 27 Intermediate Node Level 4 (IL4) Payload 28 Intermediate
Node Level 5 (IL5) Payload
[0099] The definitions and descriptions of different fields of data
packet in the payload part of the data packet are explained
below.
[0100] Temp and Light (4 bytes): These two fields constitute the
data field jointly. The Temp field contains the sensed value of the
ambient temperature and the light field contains the sensed reading
of light intensity.
[0101] Hop Count (HC) (1 byte): This field serves the purpose of
indicating the hop count of the packet at a particular intermediate
node. The hop count enables to easily and quickly fill in the next
node ID in the packet without reading the complete packet. For
example, if the hop count is 2, this means the packet has already
traveled two hops, so the next address is placed at IID-3 without
reading the complete address path.
[0102] Cluster ID (CID) (1 byte): This field in the packet
identifies the cluster ID of the sending node. Each cluster is has
a unique ID assigned by its first CH at the time of the birth of
the new cluster and all the nodes belonging to that cluster are
identified by the cluster ID.
[0103] Sensing Node ID (NID) (1 byte): This field carries the ID of
the sensing node. This unique ID is assigned to each node at
boot-up and is hard coded inside every node. Each node is given a
unique ID that is not repeated in the remainder of the network.
[0104] Intermediate Node ID (IID1-IID5) (5 bytes): These 5 fields
are used to convey the complete route information the packet
traverses from the sensing node to the sink. In the current
implementation, we have programmed the system to operate for a
maximum of six hops in including the hop from the sensing node to
the CH. If the packet reaches the sink with less than six hops, the
remaining corresponding fields are set to zero; otherwise, the
fields have the node IDs of all the intermediate heads that relay
the data packet.
[0105] Sensing Node Level (1 byte): This field contains the
assigned power level of the sensing node.
[0106] Intermediate Node Level 1 (IL1-IL5) (5 bytes): These fields
provide the power levels of the intermediate CHs.
[0107] With respect to scheduling control, similar to the S-MAC,
the inventive wireless sensor network with energy efficient
protocols implements a coordinated sleeping of the nodes to
conserve the power. The sink node serves as the data collection
center and also the schedule synchronization source of the network.
The sink node is connected via USB port to the Network Monitoring
Server for the onward processing of the received data and
monitoring of the network status. All the nodes in the network are
synchronized with the sink node.
[0108] Since the sink node is assumed to have unlimited amount of
power, the transmission of the beacon signal is done at the maximum
available power level of the radio. The sink node transmits a
broadcast beacon signal once every cycle, i.e., once in every 4
seconds. This beacon signal prompts all the nodes in the network to
sleep after predefined time and at the same time provides a counter
value. At the time of startup, all the nodes are randomly placed.
When the first node wakes up, it receives the beacon signal from
the sink. Upon receiving this packet from the sink, the CHs adjust
their timers such that they go to sleep 300 ms after the reception
of the beacon signal while the member nodes sleep after 100 ms of
the reception of the beacon signal. This means that the duty cycle
for the CHs is 300/4000 and the duty cycle for the member nodes is
100/4000.
[0109] The wake time for the CHs is chosen to be larger than the
member nodes because the member nodes simply need to sense the data
and pass it on to the corresponding CH. But the CHs not only have
to receive data from multiple member nodes, they also have to take
part in the data relaying and routing process. Therefore, they must
keep awake for longer time compared to the member nodes. The beacon
signal has a countdown value that is decremented in each subsequent
signal by 4, which indicates the time `left until the CH
responsibility shifts to another node. Therefore, all the new nodes
that boot up later in the network are also tuned to the
cluster-shifting schedule.
[0110] The link formation process 600 is represented in the
flowchart of FIG. 6. At the time of startup, a node wakes up at
step 602 and checks for the existence of any neighboring CHs at
step 604. This node waits for 5 seconds and step 606 is performed
in which if the node does not receive any signal from any nearby
node, then at step 610 it declares itself a CH of a new cluster,
otherwise, if the newly awakened node receives keep alive control
signal from an existing cluster head, at step 608 it measures the
signal strength and at step 612, if the signal strength is below
the threshold it proceeds to perform step 610 wherein it discards
the signal from the other cluster head and assumes that it has to
make its own cluster by declaring itself the head of that new
cluster. This decision is made depending on the threshold values of
the received signal strength indicator (RSSI) of the received keep
alive control packet. More specifically, when any wireless node
receives the signal it measures the power of the received signal by
reading RSSI register of the CC2420 Transceiver. CC2420 has this
built-in RSSI providing a digital value that can be read by the
microcontroller from the 8-bit register. If the RSSI value is less
than -5 dBm, it is considered that the new node is physically
sufficiently away from the CR that it is better to form a new
cluster instead of joining that existing cluster. This means that
the new node must form a new cluster and become the CH. The
threshold value of -5 dBm has been selected after extensive
experimentation and has been found to be optimum for a typical
indoor environment.
[0111] If the node receives control signal from an existing CH and
the signal strength is above the threshold, then at step 614 it
joins that cluster as a member and assigns itself the cluster ID of
that head. More specifically, if the RSSI of the received
keep-alive control packet is greater than -5 dam, then this new
node becomes member of that cluster and adopts the cluster ID that
it receives in the control packet. If a node receives signals from
two or more CHs it joins as a member of the cluster whose head is
nearer to it by measuring the signal strength of their keep alive
signals and comparing their power indications. In step 616, each
member follows the aforementioned scheduling control for members.
In step 620 a newly formed CH broadcasts a keep alive signal. In
step 622 the CH then follows the scheduling control applicable to
CHs. At step 618 the sensor node boot-up process terminates.
[0112] Regarding cluster formation, as other cluster heads wake up
they start following the same wake-up schedule which synchronizes
them with the timing of the network. With the course of time more
clusters can be formed as well as more member nodes can join the
clusters and population of the cluster as well as the network
grows.
[0113] The cluster ID is generated by multiplying the node ID of
the first cluster head node by 10. This scheme of generating
cluster IDs permits the support of a large number of clusters.
During the whole operation of the network, all the nodes keep this
same cluster ID. Each of the CHs create and maintain a cluster
table that has the node IDs of its members. This table is created
after receiving the keep-alive packets from the member nodes and is
used to select the next CH in the rotation. The next CH is selected
as the node that has the next lower cluster ID than the current CH.
In other words CHs are selected based on descending order of node
IDs of all the entries of the cluster tables. If a CH does not
receive a keep alive from a particular node for more than 3 cycles,
it is considered dead and is removed from the cluster table. New
entries are added and old ones are refreshed and updated as the
keep-alive messages are received.
[0114] The cluster tables are formed and maintained by the CHs only
and the member nodes don't need to maintain these tables. When a CH
becomes the member node after the cluster rotation, it flushes the
clusters table it was keeping and the new CH starts building its
own cluster table.
[0115] Every CH broadcasts a keep alive packet in each alternate
cycle (i.e. after every 8 s). The purpose of this keep alive is to
tell the immediate neighboring CH nodes about the presence of this
node and to keep its entry in the routing table entries. As a
result, each of the CHs refreshes and maintains a routing table to
select the proper route. All the member nodes within a cluster
regularly (once in three cycles, i.e. every 12 seconds) sense the
temperature and light of the environment and send the raw data to
their respective CHs soonest as per the multiple access algorithms.
In addition to that, the member nodes also send the keep-alive
message to their corresponding CHs once every cycle to help the CH
to update its duster table. It should be noted that the member
nodes send data only to their assigned CH (Unicast communication)
and not to other
[0116] Once the cluster head gets the data, it routes the data
towards the base with the help of the intermediate CHs.
[0117] The transmission power of the transceiver CC2420 radio unit
can be adjusted at different settings. For the sake of conservation
of the energy used in the transmission power, the communications
between member nodes and the CH have been programmed at lowest
transmission power (mode I) of the transceiver so that only the
concerned nodes within the cluster get the data with full
conservation of power. For inter cluster head communications, a
higher transmission power level (mode 3) is used in order to give
cluster heads more coverage as they need to communicate with other
cluster heads at further distances.
[0118] Regarding channel contention and Multiple Access Control in
the present invention's implementation, the member nodes sample the
temperature and light sensors every 12 s and send the sensed values
to their CHs. But because of the wireless medium and multiple
candidates` contention for transmission, every node must check
whether any signal is present on the wireless channel or not so
that collisions can be avoided. This is done both by physical and
virtual carrier sense as explained below.
[0119] In physical carrier sensing, the node scans the channel
physically to check the presence of any signal. If the channel is
found to be busy, the node backs off and tries again in the next
cycle. On the other hand, in virtual carrier sense, the node during
its wake time listens to the transmissions in its vicinity. When it
receives a frame destined for another node it extracts the duration
field value (containing the remaining time to finish the current
transmission) from the frame and puts it in the NAV register and
starts decrementing it. A member node takes decision to send a
control or data packet only when the NAV is decremented to zero and
at the same time the physical carrier sensing also turns up
negative. Similar procedure is followed when the packet is passed
from one head node to another head node for relaying the data
towards the sink node.
[0120] The virtual carrier sensing is performed before the physical
carrier sense. Since physical carrier sensing involves powering ON
the radio and actually listening to the channel therefore this step
consumes more power as compared to the virtual carrier sensing.
Therefore, it has been decided to carry out virtual carrier sensing
to decrease the probability of finding the channel busy and only
proceed with the physical carrier sensing to ensure that there is
no transmission in the neighborhood and the channel is free.
[0121] With respect to data transfer, once the data is sensed by
the sensing node and is available for transmission after making
sure the wireless channel is free, the sensing node sends a Unicast
RTS packet to its CH. When the RTS is received by the CH, depending
on whether the CH node is free or not, the CH responds by sending a
Unicast CTS packet to the corresponding sensing node. Upon
receiving the CTS, the node transmits the data packet to the CH.
After the successful delivery of the data the CH sends the ACK
signal to the member node to confirm that it has received the data
properly. If the intended receiver does not receive the packet for
some reason such as collision, the ACK is not received by the
member node and the contention process is repeated again. The same
procedure is followed for the data transfer from one CH to another
CH.
[0122] The flowchart in FIG. 7 shows the implemented MAC protocol
700 of the wireless sensor network. Steps of MAC protocol 700
include a data available step 702, a check to see if the channel is
free 704, a suspend transmission step 708, and a Request To Send
step 706 addressed to the cluster head (CH). A Clear To Send
received check step 710 is then performed, and the Member Node
waits until the CTS is received, At step 714 the Member Node (MN)
checks again for a free channel and waits at step 712 until the
channel is free. The MN then sends the data at step 716 whereupon
the MN subsequently returns to sleep at step 718.
[0123] To ensure that the whole message is transferred during the
wake interval' of the nodes, a check has been added to ensure that
the intended recipient node sends the CTS only if there is
sufficient time available in the wake period to complete the
transaction. If there is insufficient time to complete the
communication, the intended recipient does not send the CTS and the
sensing node contends for the channel again in the next wake
period.
[0124] Regarding Head Rotation Operation, as mentioned earlier, in
order to balance the power consumption among the nodes in a
cluster, the cluster head responsibility is rotated every 10
cycles, i.e., every 40 s. This is performed in a descending order
of the node IDs of the member nodes in the cluster table. Two
special types of control packets called the handshaking packets
(Table 4.2) are exchanged between the CH and the member nodes for
handshaking of the CH responsibility handover rotation.
[0125] Once the switch over of CH takes place, the new CH listens
to the keep alive messages of the neighboring CHs as well as the
keep alive messages of the member nodes to create routing tables
and cluster tables, respectively. This process takes time that
varies depending on the time taken by all the members and neighbor
CH to send their keep alive packets. At the same time, the new CH
also transmits its own keep alive message to alert the neighboring
clusters that now it is acting as the head of the cluster and all
the routing must be done through it instead of the old CH.
[0126] As the implementation details were not found during the
literature survey, the inventors of this wireless sensor network
with energy efficient protocols sought to devise unique methods to
implement the S-MAC protocol. In the process of doing so, a
protocol that was different from the S-MAC was created. Some of the
most important features of the MAC protocol 700 are listed
below.
[0127] 1. The implementation enables the nodes to communicate with
each other, the sleep times are synchronized, i.e., all the nodes
sleep and wakeup at the same time. These schedules are maintained
by every node, thus enabling them to know when the neighboring node
will be awake so that data can be sent to it. Advantageously, this
approach reduces the latency of the network.
[0128] 2. The network provides for the whole network having a
single schedule that is controlled by the sink node and followed by
all the nodes in the network. This is chosen to simplify the
scheduling process as well as to have a single master clock in the
network. If there are different groups of nodes trying to schedule
the periodic sleeping, it will result in large timing difference
among the nodes due to the natural different clock drifts speeds in
different nodes processors. Moreover, the synchronization process
remains robust to synchronization errors and simplifies timing
requirements at the same time. Additionally, all the exchanged
timestamps are relative rather than absolute.
[0129] 3. In the WSN, the infrastructure is cluster-based, the
member nodes communicate with their CHs only, and the data is
relayed in a multi-hop routing manner by cluster heads only.
[0130] 4. In the present WSN, the intended recipient node makes
sure that there is enough time left in the wake period for
receiving the data successfully before its sleep time and only in
that case the CTS signal is sent to the sender. This again
simplifies the whole operation as the schedules remain
undisturbed.
[0131] It should be understood that S-MAC protocol does not
restrict on using a predefined packet structure and leaves this on
the specific implementation. This provides the flexibility to use
custom-designed payload structures according to requirements of the
WSN.
[0132] The present WSN implements a routing layer protocol that
differs from LEACH and TEEN protocols in the following ways. With
respect to the inventive network topology, the sink node has a
constant power supply and therefore has no energy constraints. It
can transmit with high power to all the nodes. Thus, there is no
need for routing mechanism from the sink node to any wireless
network node. However, the wireless network nodes are assumed to be
far away from the sink node and because of their power constraints
it is not feasible to communicate directly to the sink node.
[0133] As shown in FIG. 4, the network topology is of a
clustered-type multi-hop routing similar to the LEACH protocol. The
exemplary topological model 400 comprises clusters 402a, 402b,
402c, and 402d, each having a number of nodes. In each cluster at
any given time, one node acts as a cluster head (CH) (e.g., nodes
A, B, C, and D), whereas the remaining nodes are the member nodes
in that cluster (e.g., nodes A-1, A-2, B-1, B-2, etc.). Only the
member nodes perform the data sensing and the CHs are responsible
for receiving the data from the member node and routing it reliably
and efficiently to the sink node S.
[0134] The CH is programmed to be dynamically rotated based on the
head rotation operations criteria defined above. The head
responsibility rotates among the member nodes in order to balance
the power consumption load between all member nodes equally.
[0135] Regarding the network tree and routing tables, during the
network setup phase, a logical network tree is created which
comprises of all the clusters placed at different power levels.
These levels indicate the depth of the tree relative to the sink
node. To decide the levels, the sink or base node (S) regularly
broadcasts a beacon signal. When any wireless node receives the
beacon signal, it measures the power of the received signal by
reading RSSI register of the CC2420 Transceiver. CC2420 has this
built-in RSSI providing a digital value that can be read by the
microcontroller from the 8-bit register.
[0136] Depending upon the received value of RSSI, every cluster
head node assigns itself a level. To assign the levels to
corresponding RSSI values, ranges have been defined so that an
adequate number of levels are created. The resolutions of the RSSI
have been carefully defined after a large number of experimentation
and testing and are shown in Table 5.
TABLE-US-00005 TABLE 5 Cluster Head Levels RSSI Range in dBm
Assigned Level RSSI .gtoreq. 0 1 -10 < RSSI .ltoreq. -1 2 -13
< RSSI .ltoreq. -10 3 -16 < RSSI .ltoreq. -13 4 -19 < RSSI
.ltoreq. -16 5 -22 < RSSI .ltoreq. -19 6 -25 < RSSI .ltoreq.
-22 7 -28 < RSSI .ltoreq. -25 8 -31 < RSSI .ltoreq. -28 9 -34
< RSSI .ltoreq. -31 10 RSSI .ltoreq. -34 11
[0137] It is to be noted here that if levels are assigned over
narrow intervals of RSSI, a large number of levels will be created.
This can produce situations in which different levels are assigned
to nodes that are physically close to each other. On the contrary,
if a wide interval is used, then in situations where the signal'
strength doesn't vary, majority of the cluster heads will be
associated to the same level. We have assigned this interval
distribution after rigorous testing of the network and monitoring
of the physical separation between the nodes in an indoor
environment.
[0138] With respect to formation and updating of routing tables,
when each of the cluster heads has been assigned a level, a logical
tree structure is formed. Each cluster head keeps sending a keep
alive control packet periodically in every cycle to inform the
neighboring cluster heads that it is alive. In this packet, the
level of the node is also broadcast. The neighboring nodes receive
this message and store the node ID and the corresponding level in
its routing table. In case, a node dies, and the neighboring nodes
don't receive keep alive message from that node for three
successive cycles, its neighbors update their routing tables
accordingly by assuming it is dead and flushing its entry from
their routing tables. In order to accommodate for the topology
changes, the routing tables are also updated every head rotation
process. This enables the network to include any new CH and to
remove the CHs that have died.
[0139] During routing operation, when a CH has some data to send to
the sink node it consults its routing table and selects the node
which is in an upper level (i.e. a level closer to the sink). For
example, considering the network of FIG. 4, the node C would have
the address table given in Table 2. When node C wants to send data
to the sink, it looks through the routing table and sends the
packet to the node that is one level higher in hierarchy or closer
(level 2) to the sink. Thus, node C will send data to Node B.
Assume that node B dies due to some reason, then the next closer
node (node A at level 1) will be the new recipient of data from
node C, and this process continues. One important rule in this
approach is that data is always sent to the CH that is at a
strictly lower tree level than the current CH. This prevents data
from cycling between the nodes.
TABLE-US-00006 TABLE 6 Node Levels Node ID Level A 1 B 2 D 4 S
0
[0140] A CH forwards the data to the CH higher up in the tree and
this continues until the data reaches the sink node S. If the
sending CH has another CH on the same power level as its own, the
sending CH will send only to the CH that is strictly higher in the
level (Lower level number) in the tree. If the sending CH finds two
CHs on the same level up in the tree, then it forwards its data to
the CH whose entry is the first one found in the routing table. The
protocol takes care of the changes in the topology of the network.
For example, if a new node is added anywhere in the network, it
either associates itself with one of the existing clusters or
otherwise starts its own cluster. This new node becomes part of the
routing mechanism automatically as if it has joined an existing
cluster; it takes turn to become CH and ultimately becomes part of
a particular routing chain. If that new node has started a new
cluster then that cluster will update and may improve the already
existing routes depending upon its location and existing power
levels in the network.
[0141] If a node dies, it is dynamically taken care of by updating
the cluster tables that ultimately reflects on the update in the
routing tables as well. We also expect that the clusters closer to
the sink node will exhaust their energy reserves sooner than the
clusters that are relatively further away form the sink node. In
that situation or for any other reason, if the whole cluster dies,
the routing tables at the remaining clusters adopt this change in
the routing information by updating their routing tables insuring
that the process of data delivery from the alive network
continues.
[0142] Regarding data delivery models of the network, there are two
formations of network possible: static monitoring and dynamic
monitoring. Dynamic events in most applications require periodic
reporting and consequently generate regular traffic to be routed to
the sink. Monitoring static events allows the network to work in a
reactive mode, simply generating traffic and reporting when an
incident occurs. The present network implements both
approaches.
[0143] With respect to synchronized network monitoring of dynamic
events, similar to LEACH, our main network acts in a proactive
manner and regularly senses and forwards data to the sink node. The
implemented network protocol in this part of the network takes care
of this job efficiently and very robustly. All the member nodes are
periodically sensing during their wake time and are sleep most of
their life in order to conserve their energy reserves. The sensing
nodes wakes up periodically and send the sensed data to their CH.
The purpose of making our network proactive is to enable the
regular reporting and monitoring of the elements in the
environment.
[0144] At the same time, like TEEN, the network also has the
ability to act in a reactive mode. A portion of the network may be
an unsynchronized group of nodes (UGN) that operates outside the
domain of the main network. These nodes remain in the sleep state
(transceiver OFF, CPU at low power mode 1) most of the time. At the
same time they keep sensing the environment continuously and only
wakeup to report an incident if the sensed value crosses a
pre-defined threshold. In contrast to the main network, these nodes
do not form clusters and don't follow any time schedules. They also
don't need to maintain the cluster tables, network topology and
routing tables.
[0145] Once there is an incident occurrence (threshold crossing of
the sensed value) ready to be reported, the node wakes up and
listens for the keep alive messages from the main network. Due to
the criticality of the situation that this node needs to report,
this node hands over the data to the CH of the main network from
which it receives the signal first, regardless of its distance and
location on the network. The UGN sensing node follows all the MAC
protocol procedures to contend for the channel and to deliver the
data reliably to the first CH it hears from. After successful
delivery of the incident, this node goes to its sleep mode again
and the data is relayed to the sink node by the main network CHs
according to their routing tables.
[0146] The implemented protocol differs from LEACH, which is also
based on a clustered approach. With LEACH, the nodes group
themselves into clusters and decide a cluster head that
communicates with its peers to route data, the cluster heads
changing randomly over time to balance the energy dissipation of
the nodes. In contrast, the present routing protocol rotates the
CHs periodically in a pre-defined manner. Moreover, in contrast to
LEACH forming clusters dynamically, the present protocol utilizes
the nodes to form fixed clusters that retain their membership
permanently. This not only simplifies the implementation, it also
results in savings of energy from the overhead of communication
required for making new clusters.
[0147] Additionally, the WSN employs multi-hop dynamic routing,
thus resulting in savings of transmission power required because
multi-hop routing is much more energy efficient in terms of
transmission power required to communicate between two nodes.
[0148] In contrast to TEEN protocol, in the present wireless sensor
network, the CHs are not erected, but the responsibility is rotated
among all the members evenly and periodically in order to load
balance the power consumption within the cluster.
[0149] In contrast to TEEN protocol which defines only two levels
of clusters, one among the member nodes and the second among the
CH, the wireless sensor network defines a total of eleven cluster
levels so that the network system can support large sizes of
networks with small multi-hop routed communications among the CHs,
thus saving substantial power for transmission.
[0150] Moreover, in the present wireless sensor network, UGN
thresholds are pre-programmed in the sensing nodes. The UGNs in the
network wake up only when there is some alarm that must be reported
to the sink. By pre-programming the threshold levels, the UGNs do
not need to wake up and listen to the sink node to receive
thresholds, thus conserving power.
[0151] In summary, the present protocol is based on a clustered
approach. Nodes are classified as either member nodes (MN) or
cluster heads (CH). Whenever a node wakes up in the network, it
listens for the beacon signal from the base station (BS) during the
start-up phase and determines the RSS for the beacon. Based on the
measured RSS the node selects one of the eleven distinct
pre-defined network tree levels and becomes the CH. Each CH
transmits a periodic keep alive (KA) packet to inform the other
nodes in the network about its existence.
[0152] All the other nodes that wake up afterwards measure the
received signal strength (RSS) of the beacon packet as well as the
RSS of the all the keep-alives (KAs) that they receive within the
start-up phase. If the RSS of any of the KAs is greater than -5
dBm, the new node decides that it is close to one of the already
existing CH and elects it as its parent and becomes the MN. If the
KAs are not strong enough, the node becomes a CH and selects its
level from the RSS of the beacon packet. Transmission within a
cluster is only from MN to CH and packets are not exchanged between
MNs. This clustered approach enables all the MN of any given
cluster to send their data to only their own parents (CHs). This
clustered scheme avoids including next hop nodes that are
physically close to each other as they become the members of the
same cluster. In addition, the clustered approach requires small
routing tables as each CH only needs to know its next hop CH and
not the MNs.
[0153] With respect to the hardware, the CC2420EM unit comes with
an Antenova antenna 202 that has a height H of about 3.5 inches, as
shown in FIG. 3. The Antenova antenna is designed to operate at 2.4
GHz with applications in WiFi, Bluetooth and ZigBee. The antenna
works with linear polarization, providing a peak gain of 2.2 dBm
and an 80% operating efficiency.
[0154] The wireless sensor network also employs a network
monitoring application (NMA).
[0155] As shown in the FIG. 8, the main screen 800 of the NMA
comprises a network viewing pane used to display the network
diagram. All the packets received by the sink node are decoded.
After extracting the routing information from the received packet,
the application draws the route in the network viewing space. In
addition, the corresponding values of the sensed parameters (light
intensity and temperature) are also displayed along the route.
Within the network viewing window, eleven horizontal lines are
drawn corresponding to the eleven available power levels. After
receiving the packet, the levels of all the nodes throughout the
path are determined and the node is drawn in the corresponding
space in the network viewing window.
[0156] A Data Viewing Window portion of the NMA is used to display
the light and the temperature values of the recently received
packet at the sink node along with the ID of the sensing node.
Although the sensed parameters are also being displayed along the
node on the network, due to the congested space in the network
viewing window the font size is relatively small. The received data
is displayed in the data viewing window with a larger font size for
ease of monitoring.
[0157] Application Control and Options portion of NMA provide
different control and option available for the operation of the
application. A Data Port Selector: As the server machine may have
several USB ports, the port to which the sink node is connected
must be selected for the operation of NMA. This dropdown menu lists
all the ports available on the server machine. The administrator
has to select the correct port based on the connection of the base
node.
[0158] Receive Data: This button opens the input/output port for
communications. The port that has been selected from the dropdown
list must be opened using this button to enable communication with
the sink node.
[0159] Refresh Network: This button clears or refreshes the network
viewing window and the network is drawn again from scratch as
packets are received.
[0160] Remove/Add Levels: These buttons invert the operation of
each other. By default the network viewing window contains the
power eve lines and the network is drawn around them. If viewing of
the network becomes difficult due to increased number of nodes, the
level lines can be removed by using remove levels and can be made
visible again by add levels.
[0161] Clear Unsync Nodes: In case of alarm, unsynchronized nodes
from the incident based portion of the network will start to
transmit data. Only five unsynchronized nodes have been programmed
to be displayed on the NMA simultaneously. Since there can be only
a few unsynchronized nodes (five for this prototype system), once
the administrator sees and acknowledges the alarm he can reset or
turn it off by the clear Unsync button which removes the unsync
node display once the alarm has been acknowledged.
[0162] Report: This button opens the report window where the
history of the routes for each node can be observed in the report
window.
[0163] Restart: This button restarts the complete application. The
port is closed and needs to be opened again once the restart is
complete.
[0164] Exit: This button shuts clown and exists the application
properly by deleting the unwanted history from the database and by
closing the communications port formally. For error-free exits, the
button must be utilized. FIG. 9 shows an example of a working
instance of NMA. The screen displays a visualization 900 of Node 7,
which routes data through cluster heads 11, 15, 4, and 10.
[0165] It is to be understood that the present invention is not
limited to the embodiments described above, but encompasses any and
all embodiments within the scope of the following claims.
* * * * *